Columbus State University CSU ePress
Theses and Dissertations Student Publications
12-2010
Tracking Floral Visitation Using DNA Barcodes
Ivan Ray Shoemaker Columbus State University
Follow this and additional works at: https://csuepress.columbusstate.edu/theses_dissertations
Part of the Earth Sciences Commons, and the Environmental Sciences Commons
Recommended Citation Shoemaker, Ivan Ray, "Tracking Floral Visitation Using DNA Barcodes" (2010). Theses and Dissertations. 91. https://csuepress.columbusstate.edu/theses_dissertations/91
This Thesis is brought to you for free and open access by the Student Publications at CSU ePress. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of CSU ePress.
Digitized by the Internet Archive
in 2012 with funding from LYRASIS Members and Sloan Foundation
http://archive.org/details/trackingfloralviOOshoe Columbus State University
The College of Science
The Graduate Program in Environmental Science
Tracking Floral Visitation Using DNA Barcodes
A Thesis in
Environmental Science
by
Ivan Ray Shoemaker
Submitted in Partial Fulfillment of the Requirements for the Degree of
Master of Science
December 2010 I have submitted this thesis in partial fulfillment of the requirements for the degree of Master of Science.
3\)e^Zo\o '^gi'TZ sL^^ Date Ivan Ray Shoemaker
We approve the thesis of Ivan Ray Shoemaker as presented here.
Date Kevin S. Burgess, Assistan Biology, Thesis Advisor
Date ]j&\\& A. Ballenger, Professor of Biology, Department Chair
Dec, 2. 2&o uX G- ^f\o^un^L Date Mhn A. Barone, Associate Professor of Biology ABSTRACT
Floral visitation resulting in interspecific pollen transfer (IPT) from non-native or invasive plant species can affect the reproductive fitness of native plant species through pollen allelopathy, stigma clogging, stylar clogging, and ovule (seed) discounting. The
prevalence of IPT and the importance of pollinators in mediating its impacts, however, remain poorly understood. Although most traditional methods for examining visitor movement are insufficient for determining rates of potential IPT, one promising
alternative is the use of DNA barcoding. Because floral visitors eat, collect or unknowingly obtain pollen, nectar and other floral tissues, plant DNA should be recoverable from their bodies, permitting molecular identification of pollen loads. To
assess the utility of plant DNA barcodes for tracking floral visitation, I collected 89 honeybees (Apis mellifera) and 49 bumblebees (Bombus spp.) from a disturbed forest edge in Columbus, Georgia and analyzed their pollen loads using the plastid DNA barcode region rbcL. The identities of monospecific pollen loads were determined by comparing sequences to a local plant reference library containing 22 native and 27 non- native plant species. The overall recovery of rbcL sequences from pollen loads was relatively high (41.6%). Based on local BLASTn analysis, 95% of monospecific pollen
loads were identifiable to the species level. Rates of heterospecific visitation were quite
high (77%), as indicated by pure heterospecific (15%) and mixed pollen loads (62%).
Collectively, these results indicate that 1) plant DNA barcode sequences can be recovered
from pollen loads; 2) species-level identification of pollen loads can be determined with
high accuracy; and 3) levels of heterospecific visitation and potential IPT can be assessed with DNA barcoding. As a new means of rapidly and effectively detecting potential pollen transfer between plant species that share floral visitors, my study demonstrates that
DNA barcode analysis of pollen loads will contribute greatly to the rapidly growing field of ecological barcoding. 1
TABLE OF CONTENTS
ABSTRACT iii TABLE OF CONTENTS v LIST OF TABLES vi
LIST OF FIGURES ....vii
INTRODUCTION 1 MATERIALS AND METHODS 4 Study Site 4 Sampling 4 Plants - 4 Insect visitors and pollen loads 5 Pollen analysis 6 Microscopy 6 DNA extraction 7 Amplification and sequencing 8 Sequence Analysis 9 Statistics 10 RESULTS 10 Sampling 10 Plant reference library 10
Insects and pollen loads 1 Pollen Load Sequences 11 Sequence recoverability 11 Validity and accuracy 12 Heterospecific Visitation 12 DISCUSSION 13 Sequence recovery 14 Species Identification of Monospecific Pollen Loads 15 Heterospecific Visitation 16 Heterospecific Pollen Loads 17 Implications 19 APPENDIX A 34 APPENDIX B 38 APPENDIX C 41 LIST OF TABLES
Table Page
1 Distribution of floral visitors collected from a disturbed forest edge on the Columbus State University campus, Columbus Georgia (N 26
32.4996° W 84.9367°) between May 1 1 and June 1 , 2009
2 Polymerase chain reaction (PCR) and sequencing success of the plant barcode marker rbcL for DNA isolated from the pollen loads of two - bee taxa; results for each visitor type and combined 27
3 Percent of pollen load sequences indicating conspecific, heterospecific or mixed pollen loads for six plant species; results for each visitor type and combined 28 LIST OF FIGURES
Figure Page
1 Selection of rbcL chromatographs (bp 1 10-1 14) for PCR amplicon produced by amplifying mixtures of plant DNA (Stachys florida and Oxalis rubra) at various ratios. Percentages indicate proportion of PCR template composed of S. florida DNA 29
2 Plant sources of sequences recovered from honeybee and bumblebee pollen loads. Numbers in parentheses indicate number of visitors collected from each individual plant species 30
3 Pollen slides and corresponding sequence chromatographs generated
from pollen loads of Trifolium repens visitors. A) T. repens, plant
reference, CSU016-1; B) T. repens, conspecific visitation, A559; C) Asteraceae (Hieracium piloselloides or Sonchus asper), identifiable heterospecific visitation, A298; D) mixed pollen load, unidentifiable
heterospecific visitation, A6 1 6 31
4 Rates of heterospecific visitation expressed as percentage of pollen load sequences yielding identifiable heterospecific or mixed pollen loads. A) Total heterospecific visitation experienced by Hieracium piloselloides, Oxalis rubra, Stachys floridana, Triadica sebifera, Trifolium repens and Verbena brasiliensis; B) heterospecific visitation by honeybees; C) heterospecific visitation by bumble bees. Numbers above bars indicate number of pollen load sequences per plant species. 32
5 Mean (±1 S.E.) proportions of heterospecific visitation. A) No significant difference was found between rates of heterospecific vistation by honeybee (n=25) and bumblebee (n=27) visitors (1-way ANOVA, F,,4=2.78, P=0.171); B) Difference between heterospecific visitation received by native (n=2) and introduced (n=4) plant species was also found to be non-significant (1-way ANOVA, F] ,9=3 .46, 33 P=0.096) .' ACKNOWLEDGEMENTS
I would first like to thank Dr. Bob Boyd, Dr. Kevin Fielman and Dr. Scott Santos from Auburn University for encouraging me to be innovative in my research and for helping me to develop the original ideas that led to the formation of this project.
I would like to thank Dr. Kevin Burgess for taking me on as a student with so little notice and with so much already on his hands. Thank you for allowing me the opportunity to contribute to this awesome field of research and for working so hard to make sure that financial support was available to complete this work.
To Dr. John Barone and Dr. Julie Ballenger, thank you for your time and constructive criticism while serving as my thesis committee members. I would like to express my gratitude to Dr. Roger Brown and Dr. Schwartz for their help in creating and photographing pollen slides. Also, thank you to so many other faculty members for providing books, equipment and overall support.
I am grateful to Dr. Bill Frazier and the Environmental Science Department of Columbus State University for providing me the opportunity to present my research at international meetings for the past two years.
Last, but certainly not least, I would like to thank Katharine Korunes, who probably deserves a thesis of her own for her contributions to this project. Without your countless hours of hard work and companionship in the field and lab, this research would have been impossible. Thank you for being so dedicated and for not running in the other direction
when I told you we would be catching bees with tweezers. 1
INTRODUCTION
The consequences of direct interactions by non-native or invasive plants with native plant species have been well-studied. In general, invasive species are considered superior competitors and may impact populations of native plant species through competition for
nutrients, water, light, and space (Brown et ah 2002). Little is known, however, about the indirect effects of non-natives, such as their impact on the pollination and reproduction of
native plant species (Brown et ah 2002, Bartomeus et ah 2008). It is now recognized that non-native plants have the ability to affect native plant species through shared pollinators
(Morales & Traveset 2009) and that the presence of non-native species can reduce the reproductive fitness of native species through alterations in visitation rates (Tscheulin et ah 2009), as well as through changes in the composition of pollen carried by floral visitors (Brown & Mitchell 2001, Nielsen et ah 2008, Kandori et ah 2009).
When a pollinator or visitor forages on multiple plant species, pollen from a heterospecific plant species may be deposited on a plant's stigma, resulting in
interspecific pollen transfer, or IPT (Waser 1978). IPT between closely related species
can result in gene flow and has the potential to significantly impact the evolutionary
trajectories of species and populations (Brown & Mitchell 2001). Some of the genetic
consequences that may occur include the reduction of adaptation to local conditions,
heterosis, or even local extinction by genetic assimilation (Ellstrand 1992). IPT can also
affect the reproductive fitness of native plant species through pollen allelopathy, stigma
clogging, stylar clogging, and ovule (seed) discounting (Brown et ah 2002, Kandori et ah
2009). In addition to the deposition of foreign pollen, IPT is often associated with a 2 decrease in the deposition of conspecific pollen, or conspecific pollen loss, which can lead to pollen limitation and reduced seed set (Morales & Traveset 2009). The prevalence
of IPT and the importance of pollinators in mediating its impacts, however, remain
poorly understood (Brown et al. 2002, Bartomeus et al. 2008).
Because most plants are animal-pollinated (National Research Council 2007), it is
inevitable that floral visitors play an important role in determining the extent to which
plant species are affected by the consequences of IPT. Thus, an examination" of visitor
pollen loads can be used to estimate IPT and its associated consequences at the level of
flowers, individuals, or even species (Bartomeus et al. 2008). Researchers, however, have
often failed to use direct examination of pollen loads when exploring pollination
mutualisms (Alarcon 2009). Instead, most studies have tracked individual insects or
used stationary observation of visits to individual plants. Floral visitors, though, are often
difficult or impossible to follow (Valentini et al. 2009), and the observation of floral
visitation can be time consuming and impractical (Memmott 1999), especially in species
with infrequent visitation or rare visitors (Widmer et al. 2000). Indirectly, researchers
have also used dye powders (Brown & Mitchell 2001), histochemical stains (Peakall
1989), and magnetic tags (Gary et al. 1971) to study pollinator movement, but these
methods are time-consuming and impractical for large-scale investigations.
Consequently, even direct examination of pollen loads through the identification of
pollen or pollinia has limited value for determining the details of interactions, as pollen
identification based on morphological characteristics is both difficult and imprecise
(Zhou et al. 2007). For investigating the effects of non-native species, direct examination
of pollen loads is also especially likely to be problematic because the impact of non- 3 native plants has been correlated with phylogenetic relatedness (Morales & Traveset
2009), and closely related species are more likely to possess highly similar pollen morphology.
One promising alternative in studying pollinator visitation is the use of DNA
barcoding (Hebert et al. 2003, Janzen et al. 2009), a technique which uses a short standardized genetic marker, or group of markers, to provide rapid DNA-based identification of organisms (Borisenko et al. 2009, Plant Working Group CBOL 2009).
The mitochondrial Col barcode region has been useful for discriminating animal species
and has already proven effective for documenting biodiversity (Hebert et al. 2004,
Milankov et al. 2008) and in conservation efforts (Stahls et al. 2009, Weese & Santos
2009). For plants, the plastid markers rbcL and matK have proven more effective (Plant
Working Group CBOL 2009). To date, researchers have used putative DNA barcodes to
determine animal diets from gut contents of bolh predators (Agusti et al. 2003, Barnett et
al. 2010, Dunn et al. 2010) and herbivores (Matheson et al. 2008, Navarro et al. 2010), as
well as from feces (Bradley et al. 2007). Because floral visitors eat, collect or
unknowingly obtain pollen, nectar and other floral tissues (Davis 1996), plant DNA
should be recoverable from their bodies, permitting similar identification of pollen loads
through the use of plant DNA barcodes.
Considering current unprecedented rates of alien plant invasions (Mooney &
Cleland 2001) and global reports of pollinator declines (NRC 2007), a new means of
rapidly and effectively inspecting plant-pollinator interactions is urgently needed. As
such, the overall goal of my thesis is to explore a novel use of plant DNA barcodes in
order to address several questions related to IPT: 1) Can DNA barcode sequences be 4 recovered from pollen loads?; 2) Can the composition of pollen loads found on floral visitors be determined to the species level?; and 3) What proportion of those pollen loads indicate heterospecific visitation and the potential to contribute to IPT? To address these
questions, I collected honeybees (Apis mellifera) and bumblebees [Bombus spp.) from a local invaded habitat and analyzed their pollen loads using the plastid DNA barcode region rbcL.
MATERIALS AND METHODS
Study Site
During May 2009, plants and insect visitors were sampled from a disturbed forest edge
habitat on the Columbus State University campus, Columbus, Georgia (N 32.4996° W
84.9367°). The site contained a moderate diversity of flowering plant species typical of
the Piedmont and Coastal Plain floras (Radford et al. 1968, Weakley 2008, USDA 2009),
as well as an abundance of pollinators, including honeybees and bumblebees. Due to the
site's disturbed nature, several invasive plant species had become established. The site
also included a number of groups (Asteraceae, Oenothera, Oxalis, Verbenaceae) with
closely related species, sharing similar pollen morphologies, which would present a
considerable challenge for the determination of pollen species by traditional methods.
Sampling
Plants
Vouchers were collected for all plant species in flower (excluding grasses) within the 5 study area, as well as within a 100 m radius of the boundary of the study area. All
specimens were identified to the lowest taxonomic level (Radford et al. 1968, Weakley
2008, USDA 2009), mounted on herbarium sheets, photographed, and stored at the
Columbus State University Herbarium (COLG) as barcode vouchers (Appendix A). To establish a pollen reference collection for microscopy analysis (see Methods,
Microscopy), pollen samples were also collected from each plant species in flower. In order to generate a local plant barcode library, 3-5 cm*" of leaf tissue was collected from each species and stored at -20°C until DNA extraction (see Methods, DNA extraction).
Insect visitors and pollen loads
Insect collections were made during four weekly collection periods from May 1 1 to June
1, 2009, between 10 A.M. and 4 P.M. During these bouts, all insect floral visitors were collected. Visitation was defined as physical contact of an insect with a flower or inflorescence combined with an associated observation of plausible foraging behavior, such as probing flowers with mouthparts, nectar robbing, or gathering pollen. Insects were collected directly from the flowers using tweezers in order to avoid contamination and once captured, were individually placed in 1.5 mL microcentrifuge tubes, stored on ice, and then frozen at -20°C.
Insects were thawed, and pollen loads were washed from each specimen by vortexing and inverting insects in Millipore (Millipore, Bedford, MA,
www.millipore.com ) filtered water for 30 s, or until a considerable proportion of the
pollen was removed. Most insects were washed with 1.0 mL EbO in 1.5 mL
microcentrifuge tubes, but larger specimens, such as Bombus and Xylocopa spp., were 6 washed with 2.0 mL H2O in 15 mL centrifuge tubes. Insects were then removed from the tubes, and the pollen load solutions were centrifuged at 17,500 rpm for 2 min. The pollen
pellet and -100 |il of the supernatant were retained for further analysis.
All insects were identified to family using Mitchell (1960); however, only honeybees and bumblebees were chosen for further analysis. The benefits of selecting
these taxa included their overall abundance at the site, agricultural importance, and major relevance to current literature, including documented pollinator declines (National
Research Council 2007). These individuals were identified to the species level using an online key (Ascher et #/.2008), pinned, and deposited at the Columbus State University
Invertebrate Museum.
Pollen analysis
Microscopy
Reference pollen material and pollen loads were examined via microscopy in order to
substantiate molecular analysis. For pollen reference material, whole flowers or anthers
collected from blooming plant species were vortexed in -1.0 mL of Millipore water in
1.5 mL microcentrifuge tubes to release pollen. Non-pollen plant debris was removed
with tweezers, and the solutions were centrifuged at 17,500 rpm for 2 min. The
supernatant was removed in order to eliminate any residual debris. For pollen load
microscopy, the pollen pellets created by centrifuging (see Methods, Insect Visitors and
Pollen Loads) were re-suspended in the remaining supernatant, and -30 uL of the
resulting pollen load solutions were processed by methods modified from Kearns & 7
Inouye (1993). Both pollen reference and pollen load solutions were diluted to 1.0 mL in their 1.5 mL microcentrifuge tubes, and the pollen was dyed via the addition of -10 (aL
fuchsin red stain. After 5 min, the solutions were centrifuged 1 min at 17,400 rpm. The supernatant was removed, 1.0 mL 100% ethanol was added, and the solutions were vortexed to eliminate any clumping. The pollen was allowed to settle and was then removed using a pipette and mixed into glycerin jelly on heated slides. The mixtures were covered with slide covers and were sealed with nail polish after cooling. The pollen reference slides were then viewed under compound microscopes and used to create a local pollen atlas. Online sources were used to provide pollen morphologies for any taxa not represented by a pollen reference slide (Davis 2001). Pollen load slides were analyzed in order to determine whether bees carried conspecific, heterospecific or mixed pollen.
DNA extraction
In order to generate a local plant reference library, -100 mg frozen or -200 mg dry leaf tissue from each plant species was pulverized via FastPrep®-24 (MP Biomedicals, Solon,
OH, USA, www.mpbio.com ), and DNA was extracted using the DNEasy Plant Mini Kit
(Qiagen, Valencia, California, USA, www.qiagen.com ) or the FastDNA® Kit (MP
Biomedicals), according to the manufacturers' instructions.
DNA was also isolated from honeybee and bumblebee pollen loads for pollen
load analysis. In total, 125 pollen loads were selected from female bees, which were
collected from six plant species that received visitation by both visitor types. Those plant
species were Hieracium piloselloides (Asteraceae), Oxalis rubra (Oxalidaceae), Stachys floridana (Lamiaceae), Triadica sebifera (Euphorbiaceae), Trifolium repens (Fabaceae), and Verbena brasiliensis (Verbenaceae). For DNA isolation, the pollen pellets created by centrifuging (see Methods, Insect Visitors and Pollen Loads) were re-suspended in the remaining -100 uL supernatant by pipetting, and 70 uL of each of the resulting solutions was processed using the FastDNA® Kit, following manufacturer's instructions, except
for the addition of a second wash step using reagents supplied with the kit.
Amplification and sequencing
Polymerase chain reaction (PCR) amplification of a 607 bp region of rbcL was performed for all successfully isolated plant and pollen load DNA. The rbcL primers (rbcL F and rbcLajf634R) and general PCR conditions were modified from Fazekas et al. (2008).
PCR amplification was performed in 20 uL reactions, each containing 2 uL genomic
DNA template (-30 ng), 0.8 U AmpliTaq Gold Polymerase with GeneAmp 10X PCR
Buffer II (100 mM Tris-HCl pH 8.3, 500 mM KC1) and 2.5 mM MgC12 (Applied
Biosystems, Foster City, CA, www.appliedbiosvstems.com ), 0.2 mM dNTPs, 0.1 mM of each primer, and 5% D-(+)-Trehalose. Following an initial step of 5 min at 95 °C for enzyme activation and template denaturation, the PCR was performed with a touchdown
amplification program of ten touchdown cycles from 58 to 53.5°C (1 min at 95°C, 40 s at
58-53. 5°C and 1 min at 72°C), 30 additional cycles of 1 min at 95°C, 40 s at 54°C, 1 min at 72°C, and a final extension period of 5 min at 72°C. PCR was also performed on known ratios of Oxalis rubra and Stachys floridana DNA mixtures to provide a reference
for interpreting heterozygous chromatographs produced by mixed pollen loads (Fig. 1).
PCR amplification products were submitted to Functional Biosciences, Inc.
( www.functionalbio.com) to be purified using ExoSAP (Exonuclease I, Shrimp Alkaline 9
Phosphatase) and sequenced in both directions with the primers used for amplification.
All plant sequences were then submitted to the Barcode of Life Data (BOLD) Systems
( www.boldsystems.org ) and GenBank ( www.ncbi.nlm.nih.gov ).
Sequence Analysis
Sequence chromatograms were edited and assembled using CodonCode Aligner version
3.0.3 (CodonCode Corporation, Dedham, MA, www.codoncode.coin ) and aligned
manually in Se-Al version 2.04 ( http://tree.bio.ed.ac.uk/software/seal/). By comparison to
heterozygous reference sequences (Fig. 1 ) and the surrounding sequence quality, putative
heterozygous positions were identified and scored as ambiguous bases. Sequences with
fewer than ten ambiguous positions were considered homozygous, indicating a pollen
load composed of a single pollen type. This single (homospecific) pollen type could then
be identified as a con- or heterospecific pollen load (see below). Sequences with ten or
greater were considered heterozygous (see Yuan et al. 2004 for a similar approach),
indicating mixed pollen loads. These pollen loads were interpreted as heterospecific
because they, by nature, contained at least one foreign species of pollen.
In order to determine the source of homospecific pollen loads, a local plant
sequence library was constructed in Geneious Pro 4.8.5 (Drummond et al. 2009), based
on rbcL sequences obtained from plant species flowering at the site. Homozygous pollen
load sequences were then compared to the local plant sequence library using BLASTn for
short nearly exact searches (v. 2.2.22 as a plugin in Genious Pro 4.8.5). Species
assignments were made by selecting sequence matches with the greatest % identical sites.
By comparing each pollen load's species assignment to each visitor's collection 10 information, pollen loads could be identified as conspecific or heterospecific, relative to
the plant species on which it was collected.
Heterozygous (or mixed pollen load) sequences were further analyzed to assess the validity of heterozygous peaks. This was accomplished by evaluating the proportion
of ambiguous bases occurring at informative sites. Informative sites were identified within the plant reference sequences, and a consensus sequence was created, which
represented all of the collected plant species. Each heterozygous sequence 'was then compared to the consensus sequence to determine the number and relative percentage of ambiguous bases occurring at nucleotide positions informative among plant species found
at the study site.
Statistical analysis
A Chi-squared contingency test was used to compare sequence recovery rates of
honeybees and bumblebees. One-way analysis of variance (ANOVA) was used to
compare proportions of heterospecific visitation by honeybees and bumblebees, as well
as to introduced and native plant species. Two-way ANOVA was used to test for the
interaction between plant nativity and visitor type.
RESULTS
Sampling
Plant reference library
Forty-nine plant taxa were collected, representing 24 families and 45 genera (Appendix
A). Over half of these (55.1%) were non-native, and three species were designated as 11 noxious weeds by the U.S. Forest Service (2010). The rbcL plant barcode region was
sequenced for all specimens collected. Full length (607bp), bi-directional sequences were obtained for 47 (95.9%). An all-to-all BLAST within Geneious Pro 4.8.5 revealed that 47 of the 49 plant taxa (95.9%) possessed unique rbcL sequences, allowing them to be easily discriminated from one another. These reference sequences are available on the BOLD
Systems website and GenBank (Appendix A). Pollen reference slides were completed for
35 (71.4%) of the taxa, and online data were retrieved for the rest.
Insects and pollen loads
A total of 389 visitors from three insect orders were acquired over 33 cumulative hours of
collection across the four sampling periods (Table 1). Most of the visitors were bees
(71.8%). Of these, 89 (31.2%) were honeybees (Apis melifera), and 49 (17.2%) were bumblebees (Bombus spp.), representing five species (Appendix B). The remaining non- apoid visitors (28.2%) included ten other hymenopteran specimens (2.5%), 47 beetles
(11.8%), and 47 flies (11.8%). Eighty-four honeybees and 41 bumblebees were selected for pollen load analysis in this study, representing 94.4% and 83.6% of the total number of individuals collected for each respective taxon.
Pollen Load Sequences
Sequence recoverability
Fifty-two (41.6%) rbcL sequences were recovered from the 125 pollen loads analyzed
(Table 2). Of these, 45 were full length and bi-directional. Twenty-five (48.1%) of these
sequences were from honeybees, and 27 (51.9%) were from bumblebees. The mean 12 overall recovery of sequences from pollen loads was 41.6%, with a much higher recovery
2 rate from bumblebees (65.8%) than from honeybees (29.7%; X = 5.67, df = 1, P = 0.02;
Appendix C).
Validity and accuracy
Twenty pollen loads (38.5%) yielded homozygous sequences, indicating a single
pollen type (monospecific), and all but one were identifiable to the plant sequence library
by > 99.5% identical sites (Fig. 2). Homozygous sequences also contained no more than
three ambiguous bases ( x = 0.2), compared to heterozygous sequences, which possessed
between 17 and 115 ( x = 55). For heterozygous (or mixed pollen load) sequences, the
proportion of ambiguous bases located at sites informative for the local flora was 93.4%
on average, with a range between 69.4% and 100% (Appendix C).
Pollen load sequences generally displayed expected results, when compared to
microscopy analysis (Fig. 3). Only ten pollen loads (19.2%) showed discrepancies
between molecular and microscopy results. In eight of these, DNA sequencing predicted
a homospecific pollen load (single pollen type), whereas microscopy revealed a mixed
pollen load. In two cases, molecular analysis revealed species diversity that was not
observed with microscope analysis of pollen loads. In addition, microscopy and
molecular analysis never indicated incongruent homospecific identities.
Heterospecific Visitation
According to both microscopy and molecular analysis, all plant species received
heterospecific visitation, as indicated by both heterospecific and mixed pollen loads (Fig.
4, Table 3). The majority of visits to H. piloselloides (66.7%), T. sebifera (90.9%), and V. 13
brasiliensis (90.0%) were heterospecific, and O. rubra and S. floridana received only heterospecific visits. However, only 47.1% of T. repens visitors carried heterospecific pollen loads. When considering honeybees alone, only 25.0% of T. repens visits indicated
heterospecific visitation, and T. repens was the only plant species on which bumblebees exhibited exclusive conspecific visitation. Although higher rates of heterospecific visitation were observed to native plant species, there were no statistically significant differences between rates of heterospecific visitation to native and non-native species (1-
way ANOVA, F,,4 = 2.78, P = 0.171; Fig. 5).
Heterospecific visitation was also found for most individuals of both honeybees and bumblebees. Sixty-four percent of honeybees and 88.9% of bumblebees were found
to carry heterospecific or mixed pollen loads (Fig. 5). Most visitors (61.5%) carried mixed pollen loads, and the remaining 15.4% were found to carry pollen from only a
single heterospecific plant species. Honeybees carried three heterospecific, 13 mixed and
nine conspecific pollen loads, while bumblebees carried five, nineteen and three (Table
3). However, a 1-way ANOVA revealed no significant difference between the groups
(Fi,9 = 3.46, P = 0.096; Fig. 5). It was also determined that there was no statistical
interaction between pollinator type and plant species nativity.
DISCUSSION
The primary goal of this study was to assess the utility of plant DNA barcodes for
research concerning pollination mutualisms. My results indicate: 1) that plant DNA
barcode sequences can be recovered from pollen loads; 2) that when visitors carry
monospecific pollen loads, species-level identification of pollen loads can be determined 14 with high accuracy; and 3) that levels of heterospecific visitation and potential IPT can be assessed with these techniques. To the best of my knowledge, this is the first study to utilize plant DNA barcodes for the molecular analysis of pollen loads.
Sequence recovery
Overall sequence recovery (42%) was higher than that previously reported for an analysis of herbivore gut contents (35%; Navarro et al. (2010). Differences in rates of recovery could partially be due to the fact that Navarro et al. (2010) attempted to amplify a plant
region with DNA extracted from whole insect bodies, while I extracted DNA from pollen loads alone. Although extracting DNA from the insects and their associated plant
material is attractive for efficiency reasons, the approach has limits as well. In this study, pollen loads were removed from the external surface of bees for extraction instead of
macerating whole insects because 1) pollen present in the digestive tract may not indicate
that conspecific pollen was ever carried on the outside of the insect, where it would be available for transfer to con- or heterospecific stigmas; 2) the abundance of insect DNA could hinder the recovery of pollen or plant DNA, due to saturation of DNA binding surfaces associated with the extraction process; 3) whole insect maceration would be difficult for large insects in a high-throughput setting and avoiding this technique will ensure a more consistent analysis across all pollinator types; and 4) whole insect
maceration prevents accession of insect vouchers.
Furthermore, the size of the organism being studied may be correlated with the
amount of plant tissue available within or on the external surface of specimens, affecting
the likelihood of recovering sufficient amounts of plant DNA. Navarro et al. (2010) 15 examined weevils, a group of beetles noted for their diminutive size (<6mm). When
considering the size of their specimens, the overall sequence recovery rate of 35% is somewhat impressive. In this study, pollen loads from Bombus spp. (~14mm) were most likely to yield high quality sequences (65.8%). Pollen loads from honeybees (~llmm), which are shorter in length and much less robust, yielded sequences only 29.7% of the time. Thus, the potential for recovering plant barcode sequences from herbivores or floral
visitors is likely greater in larger insects.
The use of the rbcL barcoding region may also have facilitated the higher recovery rates seen here, as most "barcoding" approaches to plant diet analysis have
utilized the trnL intron (Valentini et al. 2009, Jurado-Rivera et al. 2009, Pegard et al.
2009, Soininen et al. 2009, Navarro et al. 2010; but see Bradley et al. 2007, Matheson et
al. 2008). Although the trnL intron has demonstrated high species-level discrimination
and is thought to be useful for PCR amplification from degraded DNA (Kress & Erickson
2007, Taberlet et al. 2007), the rbcL region has consistently exhibited one of the highest
rates of PCR success among putative barcoding regions, second only to trnH-psbA (Kress
et al. 2005, Kress & Erickson 2007, Fazekas et al. 2008). In addition, because pollen load
analysis should not usually involve amplification from degraded DNA (except perhaps
for museum specimens), the rbcL region currently provides the most potential for future
barcoding work involving pollen loads.
Species Identification of Monospecific Pollen Loads
The construction of a local rbcL sequence library allowed much greater species-level
resolution for pollen loads than has been obtained from other analyses. In insect 16 herbivore gut content studies, Jurado-Rivera et al. (2009) and Navarro et al. (2010) compared their recovered trnL sequences to the GenBank dataset and obtained highest
matches with sequence divergences from 0-5.5% and 0-10.7%, allowing identification to
the genus-level for only 51% and 56% of sequences, respectively. Similarly, Pegard et al.
(2009) and Soininen et al. (2009), using trnL for the analyses of mammal stomach
contents and feces, were only able to identify 51.1% and 75% of species sequences to the
genus-level. When compared to a local sequence library, however, rbcL sequences from
monospecific pollen loads (this study) yielded sequences divergences of <0.5%, except
for one sequence. Likewise, Valentini et al. (2009) generated a local trnL library and
were able to identify 75% of samples to the species-level. In general, until barcode
libraries mature, the generation of local sequence databases is essential to reaching
species-level discrimination for samples of interest (Jurado-Rivera et al. 2009). My study
contributes directly to this effort and highlights the importance of developing local
barcode libraries for ecological barcoding.
Heterospecific Visitation
The high rates of heterospecific visitation observed here are similar to those observed by
other studies. For example, Bartomeus et al. (2008) reported that 77% of visitors to five
native species in Spain carried pollen from a co-flowering invasive species. Furthermore,
in a meta-analysis of sixteen studies, Morales & Traveset (2008) reported that pollinator
sharing by plant species ranged from 5 to 100% and that plant-to-plant transitions by
visitors varied between 9 and 65%, both indicating high potential for JPT. In the few
studies that have reported actual IPT, researchers have observed rates ranging from 4 to 17
50% (McLernon et al. 1996). However, actual rates of heterospecific visitation and IPT, though, vary significantly depending on many factors, including the plant species examined, other plant species present, and the species and abundances of visitors present.
The impact of IPT from invasive plants to native species is also highly variable, and the consequence of non-native invasion must be assessed on a case-by-case basis.
The contributions of individual visitor species may also depend on various conditions. For example, although bumblebee species are generally thought to be among the most constant of pollinators (Morales & Traveset 2008), my study revealed a slightly
higher rate of constancy by honeybees than Bombus spp. This finding supports evidence
that bumblebees have the ability to remain constant on several species simultaneously
without experiencing reductions in foraging efficiency (Raine & Chittka 2007). This type
of multi-species foraging is thought to be especially likely to occur when patches of co-
flowering species exist in close proximity (Raine & Chittka 2007), as was certainly case
in the collection area for this study.
Heterospecific Pollen Loads
The main limitation of pollen load barcoding is that when pollen loads are mixed,
standard Sanger sequencing methods are unable to resolve the identity of the constituent
species. I explored the use of alternative base-calling techniques to identify the major
component of heterozygous sequences. After base-calling all heterozygous positions as
the most dominant nucleotide, I was able to improve the average BLASTn percent
identity of heterozygous sequences from 90.7% to 98.4%; however, only 14 (43.8%) of
the 32 heterozygous sequences scored BLASTn hits > 99.5% and minor pollen species 18 were not identified. While this approach does provide some insight into the composition
of pollen loads, the level of neglected information is not ideal, especially for studies with small sample sizes or when rare plant or visitor species are involved (Alarcon 2009).
Several methods, though, can be used to resolve environmental samples. Vector
cloning is one method by which to isolate individual sequences from complex mixtures
(Jurado-Rivera et al. 2009), but the method can be time consuming and expensive.
Analysis of terminal restriction fragment length polymorphisms (T-RFLP) has also been
used to gain insight into complex samples. It is commonly used to assess the microbial
diversity of soils (Kirk et al. 2004) and has also been used to assess the diversity of
bacterial symbionts in termite guts (Trakulnaleamsai et al. 2004), to identify bacterial
pathogens (Nilsson & Stom 2002), and to determine the source of bloodmeals in
mosquitoes (Meece et al. 2005). However, because multiple species may produce similar
fragment lengths and because even single nucleotide polymorphisms can drastically
change a specimen's fragment profile, T-RFLP fails to provide unambiguous species-
level identification or to account for intraspecific variation.
Alternatively, pyrosequencing has been used in the identification of herbivore
diets and is capable of generating thousands of sequences per mixed sample (Pegard et al.
2009, Soininen et al. 2009). This type of analysis, which is now being provided by many
sequencing facilities, requires little if any additional equipment or skills other than those
needed for standard PCR amplification, and although previously cost-prohibitive, the
technology is becoming increasingly more affordable (Pegard et al. 2009). New advances
in pyrosequencing technology are also permitting the sequencing of fragments longer
than has previously been possible (454 Life Sciences, www.454.com ). As such, 19 pyrosequencing appears to be the emerging method of choice for ecological barcoding
and for molecular analysis of pollen loads (Mike Wilkinson, personal communication).
Implications
Overall, DNA barcoding of pollen loads offers a viable alternative to other currently
available techniques for determining plant-pollinator interactions. Most pollination
networks, for example, have been based on observation of visitation (Alarcon 2009);
however, separate observations of visitation to heterospecific plant species does not
necessarily denote interspecific visitation. In addition, even confirmed interspecific
visitation does not equate to IPT, and few studies address this by testing for the presence
of pollen on visitors or stigmas. Reliance on observation also tends to lead to assumptions
of specialization by rare plants or visitors due to an overall lack of data (Alarcon 2009).
DNA barcoding eliminates both of these problems by indicating visitation only to plant
species where a visitor has obtained significant quantities of plant tissue(s) and by
potentially providing information for visits in addition to those observed during visitor
collection. Morphological analysis of pollen loads on stigmas may be useful as a means
of assessing IPT, but it does not provide any information about the visitor(s), and the
pollen is not always easily identifiable (Zhou et al. 2007). The barcoding of pollen
present on stigmatic surfaces still remains to be explored, and as such, a combination of
pollen load barcoding and morphological stigma analysis may currently represent the
ideal approach for future IPT studies. Ultimately, pollen barcoding will be especially
useful in studies of plant hybridization or pollinator competition among closely related
plant species, where pollen types are otherwise difficult or impossible to distinguish. As 2(5 pyrosequencing improves and becomes more affordable, and as DNA barcode sequence databases mature, pollen barcoding will also become an increasingly attractive alternative. 21
LIST OF REFERENCES
Ascher JS, Eardley C, Griswold T et al. (2008) Discover Life Apoidea species guide. http://www.discoverlife.org
Agusti N, Shayler SP, Harwood JD et al. (2003) Collembola as alternative prey sustaining spiders in arable ecosystems: prey detection within predators using molecular markers. Molecular Ecology, 12, 3467-3475.
Alarcon, R (2010) Congruence between visitation and pollen-transport networks in a California plant-pollinator community. Oikos, 119, 35^44.
Barnett A, Redd KS, Frushera SD et al. (2010) Non-lethal method to obtain stomach samples from a large marine predator and the use of DNA analysis to improve dietary information. Journal of Experimental Marine Biology and Ecology, 393, 188-192.
Bartomeus I, Bosch J, Montserrat V (2008) High invasive pollen transfer, yet low deposition on native stigmas in a Carpobrotus-'mvadtd community. Annals of Botany, 102,417-424.
Borisenko AV, Sones JE, Hebert PDN (2009) The front-end logistics of DNA barcoding:
challenges and prospects. Molecular Ecology and Resources, 9(Suppl. 1), 27-34.
Bradley BJ, Stiller M, Doran-Sheehy DM et al. (2007) Plant DNA sequences from feces: potential means for assessing diets of wild primates. American Journal of Primatology, 69, 699-705.
Brown BJ, Mitchell RJ (2001) Competition for pollination: effects of pollen of an invasive plant on seed set of a native congener. Oecologia, 129, 43-49.
Brown BJ, Mitchell RJ, Graham SA (2002) Competition for pollination between an invasive species (Purple Loosestrife) and a native congener. Ecology, 83, 2328- 2336.
CBOL Plant Working Group: Hollingsworth PM, Forrest LL, Spouge JL et al. (2009). A DNA barcode for land plants. Proceedings of the National Academy of Sciences, USA, 106, 12794-12797.
Davis AR (1996) Further miscellaneous constituents of corbicular pollen loads from Apis mellifera: petals, stamens, anther threads, and fhrips. Acta Horticulturae, 437, 199-206.
Davis OK (2001) University of Arizona Catalog of Internet Pollen and Spore Images. http://www.geo.arizona.edu/palynology/polonweb.html 22
Dmmmond AJ, Ashton B, Cheung M et al. (2009) GENEIOUS v4.8. Available at http://www.geneious.coni/
Dunn MR, Szabo A, McVeagh MS, Smith PJ.(2010) The diet of deepwater sharks and the benefits of using DNA identification of prey. Deep Sea Research Part I: Oceanographic Research Papers, 57, 923-930.
Ellstrand, NC (1992) Gene flow by pollen: implications for plant conservation genetics. Oikos, 63, 77-86.
Fazekas AJ, Burgess KS, Kesanakurti PR et al. (2008) Multiple multilocus DNA barcodes from the plastid genome discriminate plant species equally well. PLoS
ONE, 3, e2802.
Gary NE, Witherell PC, Lorenzen K, Marston JM (1971) The interfield distribution of honey bees foraging on carrots, onions, and safflower. Environmental
Entomology, 6, 637-640.
Hebert PDN, Ratnasingham S, de Waard JR (2003) Barcoding animal life: cytochrome c
oxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society of London B: Biological Sciences (Suppl), 270, S96-S99.
Hebert PDN, Penton EH, Burns JM et al. (2004) Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator. Proceedings of the National Academy of Sciences, 101, 14812-14817.
Janzen DH, Hallwachs W, Blandin P et al. (2009) Molecular Ecology and Resources,
9(Supp. 1), 27-34.
Jurado-Rivera JA, Vogler AP, Reid CAM et al. (2009) DNA barcoding insect-host plant associations. Proceedings of the Royal Society B: Biological Sciences, 276, 639- 648.
Kandori I, Toshihiro Hirao T, Matsunaga S, Kurosaki T (2009) An invasive dandelion unilaterally reduces the reproduction of a native congener through competition for pollination. Oecologia, 159, 559-569.
Kearns CA, Inouye DW (1993) Techniques for Pollination Biologists. University Press of Colorado, Niwot, Colorado.
Kirk JL, Beaudette LA, Hart M et al. (2004) Methods of studying soil microbial diversity. Journal of Microbiological Methods, 58, 169-188. 23
Kress WJ, Wurdack KJ, Zimmer EA et al. (2005) Use of DNA barcodes to identify flowering plants. Proceedings of the National Academy of Sciences, USA, 102, 8369-8374.
Kress WJ, Erickson DL (2007) A two-locus global DNA barcode for land plants: the coding rbcL gene complements the non-coding trnH-psbA spacer region. PLoS ONE, 6, e508.
Matheson CD, Muller GC. Junnila A et al. (2008) A PCR method for detection of plant meals from the guts of insects. Organisms Diversity & Evolution, 7, 294-303.
McLernon SM, Murphy, SD, Aarssen, LW (1996) Heterospecific pollen transfer.. between sympatric species in a midsuccessional old-field community. American Journal of Botany, 89, 1168-1174.
Meece JK, Reynolds CE, Stockwell PJ, Jenson TA et al. (2005) Identification of mosquito bloodmeal source by terminal restriction fragment length polymorphism profile analysis of the cytochrome B gene. Journal of Medical Entomology, 42, 657-667.
Memmolt J (1999) The structure of a plant-pollinator food web. Ecology Letters, 2, 276- 280.
Milankov V, Stahls G, Stamenkovic J, Vujic A (2008) Genetic diversity of populations of Merodon aureus and M. cinereus species complexes (Diptera, Syrphidae): integrative taxonomy and implications for conservation priorities on the Balkan
Peninsula. Conservation Genetics, 9, 1125-1137.
Mitchell TB (1960) Bees of the eastern United States. North Carolina Agricultural
Experiment Station Technical Bulletin, 1, 1-538.
Mooney HA, Cleland EE (2001) The evolutionary impact of invasive species. Proceedings of the National Academy of Sciences, USA, 98, 5446 -5451.
Morales CL, Traveset A (2008) Interspecific pollen transfer: magnitude, prevalence and consequences for plant fitness. Critical Reviews in Plant Sciences, 27, 221-238.
Morales CL, Traveset A (2009) A meta-analysis of impacts of alien vs. native plants on pollinator visitation and reproductive success of co-flowering native plants. Ecology Letters, 12, 716-728.
National Research Council (U.S.). Committee on the Status of Pollinators in North America (2007) Status of Pollinators in North America. National Academies Press, Washington, DC. 24
Navarro SP, Jurado-Rivera JA, Gomez-Zurita J et al. (2010) DNA profiling of host-
herbivore interactions in tropical forests. Ecological Entomology, 35 (Suppl. 1), 18-32.
Nielsen C, Heimes C, Kollmann J (2008) Little evidence for negative effects of an invasive alien plant on pollinator services. Biological Invasions, 10, 1353-1363.
Nilsson WB, Strom MS (2002) Detection and identification of bacterial pathogens offish in kidney tissue using terminal restriction fragment length polymorphism (T- RFLP) analysis of 16S rRNA genes. Diseases of aquatic organisms, 48, 175-185.
Peakall R (1989) A new technique for monitoring pollen flow in orchids. Oecologia, 79, 361-365.
Pegard A, Miquel C, Valentini A et al. (2009) Universal DNA-based methods for assessing the diet of grazing livestock and wildlife from feces. Journal of Agricultural and Food Chemistry, 57, 5700-5706.
Radford AE, Ahles HE, Bell CR (1968) Manual of the Vascular Flora of the Carolinas. The University of North Carolina Press, Chapel Hill, North Carolina.
Rainc NE, Chittka L (2007) Comparison of flower constancy and foraging performance in three bumblebee species (Hymenoptera: Apidae: Bomhus). Entomologia Generalis, 28,081-089.
Soininen EM, Valentini A, Coissac E et al. (2009) Analysing diet of small herbivores: the efficiency of DNA barcoding coupled with high-throughput pyrosequencing for
deciphering the composition of complex plant mixtures. Frontiers in Zoology, 6, 16.
Stahls G, Vujic A, Vujic A, Perez-Banon C et al. (2009) COI barcodes for identification of Merodon hoverflies (Diptera, Syrphidae) of Lesvos Island, Greece. Molecular Ecology Resources, 9, 1431-1438.
Taberlet P, Coissac E, Pompanon F et al. (2007) Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Research, 35, el4.
Trakulnaleamsai S, Yuichi Hongoh Y, Deevong P, Noparatnaraporn N (2004) Phylogenetic diversity of bacterial symbionts in the guts of wood-feeding termites. Kasetsart Journal (Natural Science) 38, 45-51.
Tscheulin T, Petanidou T, Potts SG, Settele J (2009) The impact of Solatium elaeagnifolium, an invasive plant in the Mediterranean, on the flower visitation and seed set of the native co-flowering species Glaucium flavum. Plant Ecology, 205, 77-85. 25
USDA, NCRS. 2009. The PLANTS Database (http://plants.usda.gov, 8 September 2009). National Plant Data Center, Baton Rouge, LA 70874-4490 USA.
USFS. 2010. FEIS Invasive Plants List (http://www.fs.fed.us/database/feis/plants/weeoV index.html, 19 July 2010).
Valentini A, Miquel C, Nawaz MA et al. (2009) New perspectives in diet analysis based on DNA barcoding and parallel pyrosequencing: the IrnL approach. Molecular
Ecology Resources, 24, 1 10-1 17.
Waser NM (1978) Interspecific pollen transfer and competition between co-occurring plant species. Oecologia, 36, 223-236.
Weakley A (2008) Flora of the Carolinas, Virginia, and Georgia and Surrounding Areas, http://www.herbarium.unc.edu/flora.htm
Weese DA, Santos SR (2009) Genetic identification of source populations for an aquarium-traded invertebrate. Animal Conservation, 12, 13-19.
Widmer A, Cozzolino S, Pellegrino G et al. (2000) Molecular analysis of orchid pollinaria and pollinaria-remains found on insects. Molecular Ecology, 9, 191 1— 1914.
Yuan, CC, Li, PX, Wang, YF, Shi, S H (2004) The confirmation of putative natural hybrid species Meconopsis x cookei G. Taylor (Papaveraceae) based on nuclear ribosomal DNA ITS region sequence. Yi Chuan Xue Bao, 31, 901-7.
Zhou L, Pei K, Zhou B, Ma K (2007) A molecular approach to species identification of Chenopodiaceae pollen grains in surface soil. American Journal of Botany, 94, 477-481. 26
Table 1. Distribution of floral visitors collected from a disturbed forest edge on the
Columbus State University campus, Columbus Georgia (N 32.4996° W 84.9367°)
between May 1 1 and June 1, 2009.
Taxa Number Collected
Coleoptera 47
Diptera 47
Hymenoptera 295
bees 2X5
Apis mellifera 89
Bomhus spp. 49
other 147
non-apoid 10
TOTAL 389 ^ ^ ^ ^ ^ ! co r-- o g 2 vu ' i -st *3" OO co vo oo ^ ir; co CN CO CN
^ e? e? o_ o o o >5 CO VO OO -st oo o oo m
OO (N o\ \f co CN o\ "
^ ^ ©? i^ c5? ©5 ~ O CN co O co co \o 00 ^ ^5 CO -h oo ro O co ^ J=J ^ m o\ * CN ~ ^ £ S i -2 < \Q CN CN co „. f^. vo o\ <-* ^ r-H -S vS> 5 ^2 -2 : S-S &^ I ^ ,1"^ *> ' B ^ IH* .' - "... If i i a " & .2 S ^ • - H 5$ ul v ~ O 11111 135 O £ K K £ ; O £ Eh K. ^ vudfiiidiu sidy dds snquioq sjojisia HV 28 x> ^ o o ^ ^ ^ ^ ^ & 3 T3 100% 89% 13% 50% tN 75% 100% 56% 50% o oo h in o CI o o in o o t> t~- On co in \D 'II T3 Q , N d O 3OJ .§ d OJ S^ o (1J t/"l cm X CO ON CI o t-H oo <-h cn (N ^ ^ £ g ^ ^ ^ ^ , ^ ^ ^ ^ CI m £* —< o On K°- o O O —i tg ^ $ $ o , ^ ^ ^ - g s CO ' CO © o £ co o o o o g o if CI fa § £ II 8 ^ a d d ^ r9 O U Oh CI ' O —l \0 f-H ON o o O O CO o CO H O O H 0\ H O D s- u CO &a ?3 © _C> :. &a &a 53 <3 ea S 3 £ g 1 C 5 3 § 3 o -2 § =3 -5 ^3§ ^-w cv &a o CO ^ g H s .s g «§. d -J 11 st .3 ^ til § < < 111 § H M fill H a .si fill till O S C C ^i O 3 co E-h E^ J^ H co K E-h > H co K E^ ^ vudfiipiu sidy •dds snqiuoq sjojista HV Figure 1. Selection of rbcL chromatographs (bp 1 10-114) for PCR amplicon produced by amplifying mixtures of plant DNA (Stachys florida and Oxalis rubra) at various ratios. Percentages indicate proportion of PCR template composed of S. florida DNA. 0% £ H 59? UC^ Oh 10% C i Q 25% 50% CO 75% o o D- 90% Ou Oh 95% ~ 100% M> Figure 2. Plant sources of sequences recovered from honeybee and bumblebee pollen loads. Numbers in parentheses indicate number of visitors collected from each individual plant species. H. piloselloides (3) O. rubra (2) S. floridana (9) T. repens (17) (3 T. sebifera (11) V. brasiliensis (10) (D (3 n Brassicaceac a //. piloselloides L. sinense N. domestica Z ohiensis I sebifera E T. repens ^ brasiliensis D mixed pollen load Figure 3. Pollen slides and corresponding sequence chromatographs generated from pollen loads of Trifolium repens visitors. A) T. repens, plant reference, CSU016-1; B) T. repens, conspecific visitation, A559; C) Asteraceae (Hieracium piloselloides or Sonchus asper), identifiable heterospecific visitation, A298; D) mixed pollen load, unidentifiable heterospecific visitation, A616. A • 9 B '.'$ C homozygous (heterospecific) D heterozygous (mixed) 32 ^2 >. *ri < (%) suotjeijSjA }o uoLuodoJd 33 Figure 5. Mean (±1 S.E.) proportions of heterospecific visitation. A) No significant difference was found between rates of heterospecific vistation by honeybee (n = 25) and bumblebee (n = 27) visitors (1-way ANOVA, FM = 2.78, P = 0.171); B) Difference between heterospecific visitation received by native (n = 2) and introduced (n = 4) plant species was also found to be non-significant (1-way ANOVA, Fi.q= 3.46, P = 0.096). G 10 ° .2 P = 0.171 .1 0-80 > O 60 (L) 0.40 Oh Apis mellifera Bombus spp. Visitor Type B P = 0.096 0.80 I 0.60 - ' 0.40 - a. r C 0.20 n.nn Introduced Native Plant Nativity Status 4* -r ^D LT> o o o * c ^f •^ ui '•'1 m " u o o o o o o o o 1 o -4 3- . -r ** ' o O o o o o o o o o o o '..0 CTi 6 10 o o o o o o o o 3 3 3 3 3 3 3 3 3 3 3 3 in ''.' in in ,,, in o u u U U u u u u U U u u * z ^ - ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ o o o .-• o c * lO U3 kO * * * * * - * - * ct 3 3 3 - 3 z a fi £ S 3 £ £ £ A « £ s a a S S S z £ £ £ ^ ^ < X It 1-1 " ^ 2 W ~ " " Z Z ^ " Q ; w On 'L> OJ >y 0J OJ 1) ro in ,- " O ;i: i r '11 ' 2 t r i.'j D S CD o > < u O > 6 6 a CL g > 05 g a j E . ri 6 > | r = 6 in ; in QJ '.'.' 2 Oi -Q t* _ m -r in ^D ^ 00 o ^ tN n o o o o O o o i s o o o o O o o O O o Q ^D :jj cr> o o o i o O o o o o o o o O o o o o O O o c 3 3 3 3 3 3 3 3 3 3 3 3 in LD in IT) in l/l U in in in in 51 u u u U u u u u u u U U * cr> 00 N ^ ^ N CM «N N co n N ^ 9i -"• * c n> in CO I CO * 1 o o o o o o o o o o o o o o '-'•- .0 6 ID CO 01 6 o Q ^3" o o o o o o : o o O o o 3 3 3 -, 3 3 3 3 3 3 3 3 3 co in co CO LO L0 co L0 CO co o u U u u U U u U u U U u U u o o o o o o o o CO '1 10 kD 10 CO x x x x x x x x x x x x OJ 1 CO * * s ^ 3 S 3 3 3 i£> & & 8 2 « " £ s a S £ g a^ s & « « - ^ , X X " 2 X X 1-1 X 1-1 " X 1-1 1-1 1-1 ro ;: Cu QJ J '. 1) OJ 1. OJ cu 1 1' I u o o Q. c IT, aj (0 £ E E : u o u BC < u < LO > co a U 6 s =1 £3 CU o ^ : X OJ X ''} X _j ry j •\j u :'l = r -d V C3i J E 'h CO •S 03 f Z Qj 1 Cj. CO co CTi ,_, CTi _, in Q *t ^ CN CO ^ o ^r i 2 o O o O 2 O o o o o o o o o 3 in CO CM uo 03 un ! Q n :x o o o o o o o O o o>* o o 3 3 3 3 3 3 3 3 3 3 3 3 3 3 ' CO CO I.M ' 10 L0 L0 CO LO CO CO i £ W J O- U u u u u U u U U U u u u U *fc CO ^ cn LO LO n lO O "* m lO \r 5 o o o o o o o o o O sf ^r * in o o o o o o o o o (!) lO CO - o o o o o o o O o 10 10 *r> lO vO ID ^0 (O UD o3 fM Z> 01 LO fi s £ a n s 3 " a a A " ro ' ' & ro qi i; _ "a 2 -z. 2 2 z. to > Q- TO <: c a. ii 10 QJ QJ "3 £* a'j QJ QJ QJ 10 1 ID QJ QJ QJ m > a QJ -D TO U o O QJ 2 TO jB > ro c p :a 5 QJ a. o d Q Q U CD a: < > > g 1 TO j5 51 Q- ro u E L0 g Q L0 _J 2 i re [0 03 o O 3 c Q QJ •• "5 SJ ! Jo >& C 1 1 9- o o T3 3 x5 b w •Q QJ o .0 c P 3 a, o C '& o O O QJ u QJ C "5 6 c X3 CO o 3 &^ C t t O TO L si 1 Si S 1 -a: ^ in 0) QJ 5 o |S, E -Q *t IO i^ CO o CM ro LO UD Nali •ity Insect ID Species * * * Forage Plant Pollen Load A 301 Apis mellifera Linnaeus, 1758 I Hieracium piloselloides con A123 Apis mellifera Linnaeus, 1758 I Stachys floridana het (mix) A572 Apis mellifera Linnaeus, 1758 I Triadica sebifera het (mix) A573 Apis mellifera Linnaeus, 1758 I Triadica sebifera het (mix) A574 Apis mellifera Linnaeus, 1758 I Triadica sebifera het (mix) A5K6 Apis mellifera Linnaeus, 1758 1 Triadica sebifera het (mix) A587 Apis mellifera Linnaeus, 1758 Triadica sebifera con A589 Apis mellifera Linnaeus, 1758 Triadica sebifera het (mix) A592 Apis mellifera Linnaeus, 1758 Triadica sebifera het (mix) Ab09 Apis mellifera Linnaeus, 1758 Triadica sebifera het (mix) A628 Apis mellifera Linnaeus, 1758 Triadica sebifera het (mix) A227 Apis mellifera Linnaeus, 1758 Trifolium repens het (mix) A298 Apis mellifera Linnaeus 1758 Trifolium repens het A558 Apis mellifera Linnaeus 1758 Trifolium repens con A550 Apis mellifera Linnaeus 1758 Trifolium repens con A561 Apis mellifera Linnaeus 1758 Trifolium repens con A593 Apis mellifera Linnaeus 1758 I Trifolium repens con A605 Apis mellifera Linnaeus 1758 I Trifolium repens con A606 Apis mellifera Linnaeus 1758 I Trifolium repens con » Insect ID Species * Forage Plant A245 Apis mellifera Linnaeus, 1758 Verbena brasiliensis het (mix) A275 Apis mellifera Linnaeus, 1758 Verbena brasiliensis het A461 Apis mellifera Linnaeus, 1758 Verbena brasiliensis het A486 Apis mellifera Linnaeus, 1758 Verbena brasiliensis con A- 1 92 Apis mellifera Linnaeus, 1758 Verbena brasiliensis het (mix) A497 Apis mellifera Linnaeus, 1758 Verbena brasiliensis het (mix) A119 Stachys floridana het (mix) A133 Bombus bimaculatus Stachys floridana het (mix) Cresson, 1863 A148 Bombus bimaculatus Stachys floridana het (mix) Cresson, 1863 A310 Bombus bimaculatus Stachys floridana het Cresson, 1863 r A ,f,2 Bombus bimaculatus Trifolium repens con Cresson, 1863 A595 Bombus bimaculatus Trifolium repens con Cresson, 1863 A616 Trifolium repens het (mix) A617 Bombus bimaculatus Trifolium repens het (mix) Cresson, 1863 A493 Bombus bimaculatus Verbena brasiliensis het Cresson, 1863 A612 Bombus bimaculatus Verbena brasiliensis het (mix) Cresson, 1863 A215 Bombus fraternus (Smith, Hieracium piloselloides het (mix) 1863) A137 Bombus fraternus (Smith, Oxalis rubra het (mix) 1863) All? Bombus griseocollis (DeGeer, Stachys floridana het 1773) Ab78 Bombus griseocollis (DeGeer, Triadica sebifera het (mix) 1773) A579 Bombus griseocollis (DeGeer, Triadica sebifera het (mix) 1773) A548 Hieracium piloselloides het (mix) 40 continued Insect ID Species * Forage Plant Pollen Load A 134 Bombus impatiens Cresson, Oxalis rubra het (mix) 1863 A311 Bombus impatiens Cresson, Stachys floridana het (mix) 1863 A312 Bombus impatiens Cresson, Stachys floridana het (mix) 1863 A608 Bombus impatiens Cresson, Trifolium repens het (mix) 1863 A614 Bombus impatiens Cresson, Trifolium repens con 1863 A015 Bombus impatiens Cresson, Trifolium repens het 1863 A618 Bombus impatiens Cresson, Trifolium repens het (mix) 1863 A411 Bombus impatiens Cresson, Verbena brasiliensis het (mix) 1863 A429 Bombus impatiens Cresson, Verbena brasiliensis het 1863 A53b Bombus impatiens Cresson, Verbena brasiliensis het (mix) 1863 A41B Bombus pensylvanicus Stachys floridana het (mix) (DeGeer, 1773) * current names and authorities from Integrated Taxonomic Information System (ITIS) website, http://www.itis.gov/index.html ** determined from ITIS website; N = native; I = introduced E^E^E^E^E^E^^EE,^ E, £ E E E E E E E in in Ln u 05 00 1-1 vo X Q £ OJ u IX m u u -2 -25 -S -JS 6 E fi £ § I c a 6 S S E Q. O O O O ^ CQ CQ CQ CQ 00 O IV < < < CD 0) CU 0) -7 9 7 ^ ^ 9 ^ T 7 9 7 9 E E ^E £ E E t § £ >.j § £ £ td OJ tu 't ij qj al ai OJ cr> a) en IIIIIIIIJII CQCQCQCOCQCQCQCQ (J\ r-t tH vH E. E tt E * E m ? m ? m « E u 5 5 5 -Q « Jfl 6 E S E CQ CQ CQ CQ 00 C^ O^ (Ti in